Problem with the solver Odesolve

[vlehner]

Hi!

I'm trying to solve the deflection of a beam with a constant load along it's length.

I have a differential equation of 4th degree with four initial conditions.

The moments in the bearings are zero and the deflections too.

Now my question is: Why don't i get a solution with "Odesolve" in MP4 although i initialized all four conditions completely?

Without Units, it works!

The Error Message is:

Any help will be appreciated

Thank you.

Best Regards, Volker

[-MFra-]

Hi Volker,

In my opinion, being the differential equation of the fourth order, the initial conditions for the first and third derivatives of y are missing.

In fact, by solving analytically with Laplace, the given equation, I need all the initial conditions, as can be seen from the following calculation:

Greetings

FM

[Werner_E]

> In my opinion, being the differential equation of the fourth order, the initial conditions for the first and third derivatives of y are missing.

They aren't missing - you simply solve a different problem.

In Volkers problem we know the values of y and y'' at the end of the interval instead (something Mathcads odesolve ever so often has its problems with).

Volker already told  us and Fred had shown that the information is sufficient and odesolve is able to solve if no units are used.

So it looks like Prime cannot deal with units that perfect as we may have thought and wished.

At least I could not spot any error in the way Volker had set up the problem and the "unknown exception error" thrown may indicate a bug in Prime.

Maybe someone finds it worth the trouble to report the bug.

[-MFra-]

It could be said that mathcad automatically considers null the initial conditions missing in the solution block, which are, instead, present in the analytical solution. So there is a certain probability of error in the mathcad solution.

[Werner_E]

> It could be said that mathcad automatically considers null the initial conditions missing in the solution block,

No, it doesn't. Mathcad does not assume any specific start value. It determines the start value of the independent variable from the inititial conditions you have to provide.

[vlehner]

I will send the File with the error message and the link from my thread to PTC- let's see what they say to this.

...to be honest, I do not believe that in MP5 the problem will be solved beside many others.

[LucMeekes]

No initial conditions are missing, you just have to treat them right. The ODE is defined with:

Then the solution is (in terms of the above):

Now we have to find c0 through c3 knowing ys(0)=ys''(0)=ys(l0)=ys''(l0)=0:

That gives:

Proves thta the DE is solved, and the initial conditions are met.

Success!
Luc

[vlehner]

Impressive!

[Fred_Kohlhepp]

If it works without units, then do it without units.  (Like we did in version 15!)

[mvenich]

Symbolic solution from Maple. If you want more complicated loading conditions (like partially distributed load plus several point loads) then Maple will give the symbolic deflection/moment/shear equations (it won't handle units in the ODE, though).

[vlehner]

Thanks to all, but i need a quick solution like creating a short solve block working with units.

I want a quick numerical solution- a short solution like engineers want to have for work.

When having a beam for example with 4 forces and a moment, the symbolical solution kills time and is not practical.

Best regards, Volker

[Fred_Kohlhepp · Accepted solution]

" I need a quick solution like creating a short solve block working with units."

Attached is your quick solution with units.  Note that you can make both I and q functions of position along the beam.

Enjoy.

[vlehner]

Thanks Fred!

That hit the nails head.

I have now a short numerical solution with units at the end.

The function "UnitsOf" was new to me.

Thanks to all for your efforts.

Best regards, Volker